JPH11142231A - Noise analyzer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a noise analyzer by which the magnitude and the arrival direction of an arriving noise in terms of the sense of hearing are specified at every frequency by a method wherein, on the basis of a plurality of noise signals obtained by a noise detection means, a cross-correlation is analyzed in every frequency band. SOLUTION: A microphone 21 and a microphone 22 are attached to parts corresponding to both ears of a dummy head which simulates the head of a human being, and an arriving noise is detected so as to be stored. In an element- sound arrival-direction computing device 6, noise data trains on two time bases having a simultaneously property are FFT-ed, a spectrum is extracted in every 1/3 active band, the extracted spectrum is inverse-FFT-ed, and a waveform on a time base which is passed through every 1/3 octave BPF is found. A cross- correlation computing operation is executed to a data train in every two-channel frequency band. When the arriving noise is analyzed regarding all frequency bands which are measured inside a compartment, its arrival direction and its magnitude can be analyzed, and the generation place of an element sound having every frequency band can be specified.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise analyzer for determining the direction of arrival and the loudness of human hearing per frequency for incoming noise. In particular, measuring the noise coming to the driver's ear in a running vehicle,
This is effective for specifying the noise source in the vehicle. For example,
Calculate the noise that reaches the vehicle interior during traveling, especially the vehicle engine noise, wind noise, road surface noise, vehicle interior muffled sound, etc., based on the masking effect, which is a human auditory characteristic, and the arrival of the component sound It can be used as a noise evaluation device that specifies the direction and size and evaluates the degree of discomfort in terms of hearing from the specificity.

[0002]

2. Description of the Related Art In-vehicle noise is caused by wind noise, road surface noise, and the like, which are unpleasant to persons who are unrelated to the engine explosion cycle and the order components based on the engine explosion cycle, such as engine noise and muffled car noise. In addition, depending on the running state of the automobile, noises such as air leaking from the driver's seat window and the passenger's seat window are perceived as noise depending on the running state of the automobile.

Conventionally, as a method for measuring such noise, particularly wind noise and leak noise, Japanese Patent Laid-Open No. 3-246426 has been disclosed.
An aerodynamic noise measurement method described in Japanese Patent Application Laid-Open Publication No. H10-260, is known. The outline will be described below. A large number of microphones serving as noise detection means are mounted in a matrix on the driver's seat window, which is known to have a large effect on aerodynamic noise, that is, in a side window, and their pressure sensing surfaces are arranged flush so that they do not protrude from the glass surface. I do. Changes in pressure during traveling are simultaneously output from the large number of microphones arranged in a matrix, and are captured and stored in a computer having an A / D converter and the like. Then, among the stored pressure fluctuation data, the microphone having the highest fluctuation pressure level is set as the reference microphone, and its output is set as the reference value P0 at the reference position. The pressure data Pm and (m = 1, 2,
), The sound quality, which is the magnitude and frequency characteristics of the noise, is determined using a predetermined aerodynamic noise equation.

More specifically, FFT (high-speed data) is applied to pressure data P0 and Pm from each microphone.
(Rie transform), etc., the whole frequency band is divided into n = 1 to 30 bands, and each frequency band is a spectrum having a band structure in which Fn and the fluctuating pressure value at that time are PmFn. Express. In particular, the influence of the pressure value P0Fn of the reference microphone in the frequency band Fn on the surrounding area is described by the pressure PmFn of the surrounding microphone.
Coefficient S having the dimension of the area obtained from the correlation with
Expressed as Fn, the sound intensity X exerted on the driver's seat by a certain frequency band Fn measured by the reference microphone, that is, the loudness of the noise radiated from the side window is represented by the following equation: Yn = PFn · Fn · It is assumed that it is obtained as the sum of root mean squares of SFn. To evaluate the sound quality at that time, the values of Yn for n = 1 to 30 are compared, and if the maximum Yn or the prominent Yn exists, the frequency band Fn is assumed to be representative of the sound quality.

[0005]

It is known that a human auditory characteristic (acoustic psychological model) has a masking effect. When a large sound in a certain frequency band is given to a human auditory sense, a frequency near the same is given. Sounds below a certain level of the band are inaudible or difficult to hear. That is,
A noise signal collected by mechanical-electrical conversion of air pressure fluctuation with a capacitance type microphone or the like is different from a noise signal sensed in the peripheral and central nervous systems via the human outer, middle and inner ears Even if the noise emitted from the side window is measured and evaluated simply by adding the output of a large number of microphones and without taking into account the auditory masking effect of the person as in the conventional example, the measurement is not always audible. Did not become. For example, if the road surface noise and wind noise during high-speed driving are measured separately, both are perceived as unpleasant,
The evaluation is fairly low, but when both are heard by a person at the same time, they are masked with each other. In particular, when the volume of the wind noise and the volume of the road surface sound are 2: 1, the space is evaluated as a well-balanced space. Because

As another human auditory characteristic, when noise having a specific frequency is perceived in a specific direction,
Although the degree of discomfort is increased by judging that the noise is abnormal, the above-described method of the related art only seeks the loudness and sound quality of the noise, but does not specify the direction. In addition, other methods have also been proposed to collect noise and perform cross-correlation analysis to detect the direction of abnormal noise using a large number of microphones, all of which are based on human hearing characteristics. However, the method does not specify the direction for each frequency band, and is an inefficient measurement evaluation method or measurement apparatus.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and has as its object to specify, for each frequency band, an audible magnitude and an arrival direction for each frequency band. . Furthermore, by analyzing the noise obtained by a microphone, etc., by approaching the acoustic characteristics of the audibility perceived by humans, the noise is analyzed.
An object of the present invention is to make it possible to easily take measures against abnormal noise. Further, the present invention aims at realizing more comfortable vehicle interior sound by analyzing the incoming noise in a vehicle, and an object of the present invention is to realize a substantial noise evaluation device that reflects human perception.

[0008]

In order to achieve the above object, a noise analyzer according to claim 1 analyzes a arriving noise to determine a direction of arrival of the noise. Detecting means for detecting noise at a plurality of locations at the same time, and band level analysis which is an analysis method for converting a plurality of noise signals detected by the noise detecting means into a human audible magnitude. And a masked spectrum analysis, and calculates an audible noise level for each frequency band, and an audible noise level calculation unit, and a plurality of noise signals obtained by the noise detection unit. Element sound arrival direction calculation means for calculating the arrival direction of the element sound forming the noise by performing the cross-correlation analysis every time.

One specific method of calculating the direction of arrival for each frequency band is to obtain a cross-correlation value of signals perceived by the left and right ears extracted for each frequency band by a binaural signal processing model, There is a method of determining the direction of arrival of the sound component in the frequency band based on the relationship between the geometrical arrangement of the human binaural ears and the sound source from the delay time at which the correlation value becomes highest. At this time, the distance to the sound source may be obtained for each frequency band.
Furthermore, in a plurality of detection signals, the direction of arrival may be determined from the loudness of the audibility of each frequency band. For example, the direction of arrival can be determined by a function of the ratio of the loudness of each frequency band by a masked spectrum analysis of signals detected by the left and right ears. In this case, the direction of arrival may be determined in a high-frequency region by this method and in a low-frequency region by correlation analysis.

[0010] Further, the above-mentioned noise analysis device further includes an auditory discomfort degree based on the calculated value calculated by the elementary sound arrival direction calculating means and the calculated value obtained by the audible noise magnitude calculating means. The noise evaluation device may be configured by adding an auditory discomfort evaluation means to be evaluated. As an example of a specific process of the auditory discomfort evaluation means, the discomfort may be obtained from the degree to which the auditory loudness and arrival direction of the sound in each frequency band are determined. That is, as the human auditory characteristics, when the direction of arrival of the sound is clearly determined, the degree of discomfort is large even if the sound volume is small, and conversely, when the direction of arrival of the sound is not determined. It is known that even if the sound volume is relatively large, the degree of discomfort is small. The degree of discomfort in the sense of hearing may be obtained by using the hearing characteristics.

[0011]

According to the present invention, a plurality of incoming noises are detected by a plurality of sound pickup devices such as a plurality of microphones of the noise detecting means. These noise signals are sent to the noise level calculation means for hearing,
First, it is converted into a plurality of frequency spectra for each frequency band by band-level analysis, and furthermore, by performing a masked spectrum analysis, the magnitude of the audible noise perceived by a person, that is, the noise's masked spectrum, respectively. Is converted. The plurality of noise signals obtained by the noise detecting means are sent to the element sound arrival direction calculating means. The element sound arrival direction calculation means performs a cross-correlation analysis on each of the plurality of transmitted noise signals for each frequency band, and thereby calculates the arrival direction of the element sound forming the noise. In this way, it is possible to determine the loudness and arrival direction of the audible sound for each frequency band of the noise. Since the clarity of the arrival direction of the component sound is closely related to the recognition of the abnormal sound, it can be used for determining the degree of discomfort together with the auditory loudness of the component sound.

As shown by the above series of signal processing,
Incoming noise detected as mechanical vibrations of air by multiple microphones is converted into psychological auditory noise felt by the human auditory peripheral nervous system and central nervous system, and its directionality is calculated. And more substantial measurement and evaluation close to human sensitivity, including abnormal noise recognition.

[0013]

As described above, according to the apparatus of the present invention, the direction of arrival of the sound and the loudness of the sound perceived are determined for each frequency band, so that the degree of determination of the direction of arrival is improved. As a result, the noise source can be easily specified, and noise countermeasures can be taken. For example, it is possible to identify a location where a wind noise is generated when the vehicle is running, and to take measures against noise such as changing the shape so that the wind noise is not generated.

Further, since masked spectrum analysis is used in consideration of the masking effect of the human sense of hearing,
It is possible to realize the measurement of each elemental sound and the direction of arrival of the elemental sound with extremely high accuracy compared to the related art. As a result, the location or cause of the abnormal noise can be specified with high accuracy. In addition, since the direction analysis is performed for each frequency band, the type of sound that cannot be perceived and its direction of arrival can be clearly analyzed in terms of audibility, and as a result, a more comfortable vehicle interior space can be provided. . Further, a noise evaluation device for determining the degree of discomfort in consideration of the loudness of the auditory sound and the degree of definiteness of the direction of arrival using the device of the present invention and the unpleasant characteristics of human hearing can be configured. For example, when the direction of arrival is large, the degree of discomfort is large even if the sound volume is relatively small, and conversely, when the direction of arrival is small, the sound volume is large. It is possible to obtain a noise evaluation device using auditory characteristics such as a small degree of discomfort even if the size is relatively large.

[0015]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram illustrating an example of the present embodiment. The device of the present embodiment is a noise evaluation device using the noise analysis device of the present invention. The noise evaluation device includes a noise detection device 2, an audible noise magnitude calculation device 4, an element sound arrival direction calculation device 6, and an audibility discomfort degree calculation device 8. The noise detection device 2 receives various sounds (noise)
And a noise signal collecting device 20 having a DAT (Digital Audio Tape) for storing the collected sounds. Each of the microphones 21 and 22 is provided at a plurality of different positions in the vehicle interior. Specifically, the microphones 21 and 2 are placed on the portions corresponding to both ears of a dummy head simulating a human head.
2 is attached, and noises arriving at both ears are detected and stored.

In this embodiment, the blue color shown in FIG.
rt et al. modeled on the theory of human binaural signal processing proposed in 1983. In this model, sound waves are transmitted through the outer ear, middle ear, and cochlea to a group of finely divided hair cells, and the vibrations of these hairs are transmitted as electrical signals through the nervous system. It is assumed that sound quality, intensity, direction, pleasure, discomfort, etc. are determined by pattern recognition after integrated processing by ear cross-correlation.

In the present invention, in the calculation algorithm in the component sound arrival direction calculation device 6, the detected noise is divided into each frequency band by a critical band filter, and the signal waveforms (from the left and right microphones in each frequency band) are detected. The model is used to calculate the cross-correlation value on the time axis. The functions of the critical band filter and the half-wave rectification, which are signal processing models for the outer ear, the middle ear, the cochlea, and the hair cells, are also used for octave level analysis in the audible noise magnitude calculator 4. The auditory noise magnitude calculation device 4 further uses a human audibility characteristic that each frequency component masks another frequency component, for example, when a low frequency component is present, a high frequency component is hard to hear. . That is, the noise level calculation device 4 for the sense of hearing
Performs a masked spectrum analysis. Also,
The auditory discomfort calculating device 8 calculates the discomfort as a comprehensive evaluation of the sound loudness for each frequency band and the determinism of the arrival direction for each frequency band.

In the noise detection device 2, sounds detected by a plurality of microphones 21 and 22, which are capacitance type sensors, are digitized and sequentially converted in time series by a noise signal collection device 20 to record a DAT or the like. It is stored in the device 201 (FIG. 4). The audible noise magnitude calculation device 4, the component sound arrival direction calculation device 6, and the audibility discomfort degree calculation device 8 can be configured by hardware devices, but are realized by software by a computer device shown in FIG. Is also possible. The computer system shown in FIG. 4 includes a storage device 32 such as a hard disk device for storing a signal processing program in the device of the present embodiment and noise data stored in the recording device 201, and a CPU (CPU) for performing arithmetic processing. A central processing unit 30) and a RAM (random access memory) 31 as a work area.

The plurality of microphones 21 and 22 are specifically installed at both ear positions of the dummy head provided in the passenger seat, and detect the noise in the cabin during traveling, and as shown in FIG. The data is temporarily recorded in the recording device 201 as a digital value in time series for each of the left and right channels. The vehicle interior noise data recorded in the recording device 201 is taken into the computer system for each channel and stored in the storage device 33 or the RAM 31 described above.

Normally, noise detected in the interior of an automobile during traveling includes an order component whose fundamental frequency is the engine explosion cycle and a random noise component such as a road surface or wind noise caused by traveling, and these are mixed. ing. For example, when the number of cylinders of the engine is 6 and the number of revolutions is 1800 rpm, the explosion occurs six times in two revolutions, so that the fundamental frequency is 90 Hz. Wave components (order components) are detected. These noise signals are collected by the microphones 21 and 22 installed at the left and right ear positions.

The audible noise magnitude calculating device 4 is constituted by a CPU 30 which operates according to a numerical calculation program. In this device 4, as shown in FIG. 2A, a frequency spectrum analysis is first performed on the noise signal by FFT (Fast Fourier Transform) to show the relationship between the frequency and its intensity distribution. Although only the frequency spectrum of one channel is shown in FIG. 2A, this operation is also performed on the noise signal of the other channel.

By the way, the human auditory characteristics are logarithmic characteristics with respect to the frequency and the intensity of the sound. In performing the processing on the auditory sensation, the signal obtained above is handled over all frequencies. It is not necessary, and an octave band level analysis according to a human hearing is usually performed. Octave-band analysis is the power of the signal obtained from each band-pass filter when band-pass filters having a pass band whose high-band cutoff frequency is twice as high as the low-pass cutoff frequency are arranged so that the bands are continuous. This is a method of calculating the average value of. According to such a method, only signals that are significant in terms of audibility are extracted. In the present embodiment, in order to investigate in more detail, 1/3 octave analysis in which one band is set to 1/3 octave was employed.

Further, human hearing has another important characteristic called a masking effect as described above.
That is, in the spectral space, a spectrum, a disturbing sound (masker), and a signal sound (maskee) which are close to each other affect each other, and the one having a small spectrum intensity is not perceived or hardly perceived. is there.
The effect of the masker is centered on the interference sound frequency.
As shown in FIG. FIG.
(A) shows the relationship between the minimum audible curve with respect to the frequency of each signal sound and other spectral lines when a noise signal of 365 Hz to 455 Hz and 80 dB is input;
A signal tone (maskee) above the minimum audible curve is perceivable, and a signal below is inaudible. Although not described in detail, the minimum audible curve is not uniform but has frequency dependence and intensity dependence. As basic data of human hearing characteristics, the minimum audible curve is obtained for each center frequency of the disturbing sound and for each intensity. The curve is stored in the storage device 32.

Further, regarding the frequency of the disturbing sound (masker) in the masking phenomenon, the parameter called critical bandwidth is used.
Data exists. It is a phenomenon that the masking effect does not change even if the frequency width of the disturbing sound (masker) is gradually increased around the signal sound frequency, even if it is increased beyond a certain width. Almost match. In other words, in hearing, only components that have passed through the same band-pass filter mutually interfere (addition and subtraction). To consider the masking effect, octave band level analysis according to human hearing is effective. Is shown.

Therefore, in the actual masked spectrum analysis, within each frequency band (critical band) calculated by the 1/3 octave band level analysis, the intensity exerted by the interference frequency component (masker) is as described above. The weighted integration is performed using the minimum audible curve, and the value is subtracted from the signal sound intensity. As a specific example, a spectrum converted into a band structure is shown in FIG. If B3 is regarded as a signal tone (maskee), the signal tone B3 is B1, B2, B4
When the intensity of the noise signal integrated in the frequency band (within B3) exceeds the signal sound intensity, the signal sound cannot be perceived.
As a result of the masked spectrum analysis, the value of B3 is calculated to be zero. When the integrated value of the interference frequency component is lower than the signal sound intensity, the difference is calculated as a result of the masked spectrum analysis. Such processing is performed for each frequency band (critical band) by masked spectrum analysis.

As a result, the left and right noise signals are as shown in FIG.
The conversion is performed as shown in (b) and (c). This is,
Actually, it is the frequency spectrum of the vehicle interior noise perceived by the left and right ears. For example, when focusing on a frequency band having a strong spectrum intensity of 90 Hz to 180 Hz, the right side is a masking effect, and the spectrum intensity is almost the same. It is zero. Also, the intensity of the spectrum varies depending on the arrival direction of the noise. For example, FIG.
As shown in FIG. 5, when an elementary sound arrives at an angle formed by the wave front and a line connecting both ears with θ, the difference in the reach distance between the two ears is 2r · sin θ, so that about 2r sin θ /
While reaching the left ear with a time difference τ of v (v is a sound speed of about 340 m / sec at normal temperature), the right ear directly senses noise, while the left ear senses an indirect wave wrapped around by a direct wave. Therefore, for example, the component sound spectrum becomes small as shown in FIG. Conversely, the masked spectra of a certain frequency band (element sound) are compared with each other, and if there is a difference in intensity, it can be determined that the element sound has arrived from the one with the higher intensity. At this time, the direction can be specified by a function having a ratio of the left and right component sounds. For example, if the ratio is 1, it can be determined that the sound of the frequency band is coming from the front or the rear, and if the right component sound / left component sound is sufficiently large, it comes from the right, If the right element sound / left element sound is sufficiently low, it can be determined that the sound arrives from the left.
Also, if the right element sound / left element sound is between a sufficiently large value and 1, as the ratio increases from 1, the arrival direction changes from the front or the rear to the right. can do. Furthermore, if the right element sound / left element sound is between a sufficiently small value and 1, as the ratio decreases from 1, the arrival direction changes from the front or the rear to the left. can do. The determination of the direction of arrival based on the loudness of such a sound is effective for a sound in a frequency band higher than a certain frequency (for example, 1120 Hz).

Further, in order to further examine the arrival direction of the element sound, the element sound arrival direction calculation device 6 executes the following processing. This device 6 is also realized by software processing by a predetermined program by the CPU 30. The component sound arrival direction calculation device 6 performs FFT on noise data strings (channel 1 and channel 2) on two time axes having the same time characteristic, and extracts a spectrum for each of the above-mentioned 1/3 octave band. Then, the extracted spectrum is subjected to inverse FFT transform to obtain a waveform on the time axis that has passed through each 1/3 octave bandpass filter. The data sequence on the time axis in this frequency band b is represented by Xbi (i = 1,.
N = 2048), Ybi (i = 1,..., N = 204)
8).

The following cross-correlation calculation is performed on the data strings in each frequency band of the two channels.

Cbk = Σ (Xbi · Ybi + k) / N (added by i) (1) where b is the frequency band number, N is the number of sampling points on the time axis, and k is the phase difference number. .

From equation (1), for each b, C for k
Calculate the change characteristics of bk. Next, k at which Cbk takes the maximum value is calculated for each b. Thus, in the two signals,
The phase difference number k having the largest cross-correlation value is obtained. From the phase difference number k and the sampling period, the time difference between the two signals for each frequency band is calculated. Figure 5 shows this situation.
Shown in The curve in FIG. 5 indicates Cbk, the black dot indicates Cbk the maximum value in each frequency band, and the value on the horizontal axis at that point is the time difference.

In FIG. 5, each running speed (1) 60
miles / h, (2) 70 miles / h, (3) 80 m
The result of analyzing the noise collected for ile / h is shown. The band number b in FIG.
The frequency band corresponding to the wind noise of 6,5 is shown to be delayed by the left microphone. That is, this elemental sound comes from the right side, and the above-mentioned time difference equation 2r
The arrival direction θ is calculated from sin θ / v. Also,
Similarly, other element sounds (muffled sound, road surface sound, engine sound)
Is also calculated by a series of processes.

Further, since these component sounds usually fluctuate due to the direction of the wind or the road surface condition, the above analysis is performed at regular intervals (for example, every one second) when calculating the direction of arrival. , Its frequency needs to be determined. FIG. 6 is a graph showing the direction and magnitude of the sound obtained by analyzing the frequency band number 7 every second, and a large number of θ occur in the range of 60 ° to 90 °, It is shown that noise has arrived from that direction. As described above, by analyzing all the frequency bands measured in the vehicle cabin, the direction of arrival and the magnitude are analyzed, so that the location where the component sound having each frequency band is generated can be specified.
(FIG. 7)

It is also known that the human central nervous system senses the arrival of low-frequency noise at the above-described time difference, and senses the direction of high-frequency noise by the difference in spectrum intensity. As the arrival direction of the element sound is clear,
Auditory characteristics for recognizing abnormal sounds, that is, unpleasant sounds, are also known. According to such central nervous or auditory characteristics,
Based on the various calculated values obtained above, the auditory discomfort degree calculating device 8 evaluates each elemental sound.

That is, when the wind noise as shown in FIG. 8 is evaluated, it can be roughly divided into two processing steps. One is a process of obtaining the intensity of a high-frequency band of 1120 Hz or more in the masked spectrum of the noise obtained by the above-described auditory magnitude calculation device 4, and the other is a left-right correlation analysis and a clear direction. This is a processing process by the element sound arrival calculation device 6 including a band calculation process.
Sum of the following low-frequency band intensities (direction 30 ° to 90 °)
Is what you want. Then, the sum of the intensities obtained in both processing steps is comparable to the integrated processing on the human central nervous system, and an index indicating the clarity of the arrival direction of the final component sound, that is, an index indicating the degree of discomfort Become.

FIG. 9 shows a correlation between the evaluation value of the wind noise obtained by such processing and the sensory evaluation value obtained through a large number of subjects. It shows a very high correlation coefficient (r = 0.95), and the series of processing according to the present invention is extremely effective, and the calculated value can be used as a new noise evaluation value.

While one embodiment of the present invention has been described, various other modifications are possible. For example, in the present embodiment, the noise signals from a plurality of microphones are once taken into a recording device such as a DAT, and then offline processing is performed for analysis by a computer, but as shown in FIG.
A so-called real-time process may be adopted in which noise data from a plurality of microphones is directly taken into the A / D conversion computer and processed. Also, in the apparatus for calculating the noise level in terms of audibility, program means such as FFT (high-speed Fourier transform) including a computer was used. However, as shown in FIG. 11, an FFT analyzer 40 composed of an electronic circuit was employed. The output may be captured in a storage medium.

Also, in the octave band level analysis used in the present embodiment, the frequency band other than the octave band level was removed by a program from the noise signal taken in the storage medium. The signal may be extracted as an analog signal by a D / A converter, passed through an octave band-pass filter constituted by an electronic circuit, A / D converted, and input to a computer again. Further, in this embodiment, in order to simplify the explanation, two microphones are installed at the ear positions of the dummy head to take in a plurality of noises with a time difference and an intensity difference. A microphone may be provided to provide a noise evaluation device that can also search in the vertical direction. Although only the direction of arrival was calculated, two time differences for each frequency band of the three signals were calculated using three microphones as described above, and two spherical equations for sound wave propagation were established. By solving, it is possible to specify the position of the sound source.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a noise evaluation device according to an embodiment of the present invention.

FIG. 2 is a characteristic diagram showing a frequency spectrum calculated by an audible noise magnitude calculation device and an audible signal spectrum perceived by both ears.

FIG. 3 is an explanatory diagram showing a binaural signal processing model in which incoming noise is processed.

FIG. 4 is a block diagram when the apparatus according to the embodiment is configured by a computer system.

FIG. 5 is a characteristic diagram showing a correlation value curve of two sound waves in each frequency band and a time difference between the sound waves obtained by the noise arrival direction calculation device.

FIG. 6 is a characteristic diagram showing a direction in which noise in a certain frequency band arrives with respect to measured noise, and showing a frequency in that direction.

FIG. 7 is a characteristic diagram showing a result of obtaining a direction of arrival of noise in each frequency band and a magnitude of the measured noise.

FIG. 8 is an explanatory diagram showing a flow of processing of windbreak evaluation, which is one of the elemental sound evaluations.

FIG. 9 is a characteristic diagram showing a correlation between a value obtained by an auditory discomfort calculator and a sensory evaluation value.

FIG. 10 is an explanatory diagram illustrating a perceived signal and a non-perceived signal in masked spectrum analysis.

FIG. 11 is a block diagram showing another modification of the present invention.

[Explanation of symbols]

 Reference Signs List 2 noise detection device 4 audible noise magnitude calculation device 6 elemental sound arrival direction calculation device 8 audible discomfort calculation device 20 noise signal collection device 21, 22 microphone

Claims (1)

[Claims] 1. A noise analyzing apparatus for analyzing an incoming noise to determine a direction of arrival of the noise, comprising: a noise detecting means for detecting the incoming noise at a plurality of locations at the same time; Performs band-level analysis and masked spectrum analysis, which are analysis methods for converting a plurality of noise signals into human audible magnitudes, and calculates the audible noise magnitudes for each frequency band. Calculating the arrival direction of the component sound forming the noise by performing a cross-correlation analysis for each frequency band based on a plurality of noise signals obtained by the noise level calculating means on the audibility and the noise detecting means. A noise analysis device comprising: a component sound arrival direction calculation unit.